tools for voc or coco
Project description
图像处理工具包
- 处理 COCO, VOC 格式的数据集
- 包含数据集互转、切分、检查
- 对coco格式数据进行评估(基于漏检率、错检率)
数据集格式说明
VOC
VOC
|
|---------Annotations
|-------xxx.xml
|---------ImageSets
|
|-------Main
|-------trainval.txt
|---------JPEGImages
|-------xxx.jpg
|---------labels.txt
# ImageSets 非必须
COCO
COCO
|
|---------Annotations
|-------train.json
|---------train
|-------xxx.jpg
使用方法
- pip 安装
pip install cspdataset
使用方法
- pip 安装
pip install cspdataset
- 命令行式使用
datatools COMMAND
- 参数说明
usage: datatools.py [-h] --form {coco,COCO,VOC,voc} --data_dir DATA_DIR
[--split]
[--division_ratio DIVISION_RATIO [DIVISION_RATIO ...]]
[--transform] [--output_transform OUTPUT_TRANSFORM]
[--check] [--output_check OUTPUT_CHECK]
[--labelfile LABELFILE] [--eva EVA]
tools for voc/coco dataset
optional arguments:
-h, --help show this help message and exit
--form {coco,COCO,VOC,voc} the form of the dataset
--data_dir DATA_DIR the dir of the dataset
--split Set a switch to split
--division_ratio DIVISION_RATIO [DIVISION_RATIO ...] data segmentation ratio
--transform Set a switch to transform
--output_transform OUTPUT_TRANSFORM dir for output
--check Set a switch to check
--output_check OUTPUT_CHECK dir for output
--labelfile LABELFILE the labels.txt from the platform
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for cspdataset-0.4.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 94464c16dbe5df1b00406e69576eda2d19f21a6a500f197f06fe7ebed77a0cb3 |
|
MD5 | 640c7cfda51143e21cb4442ad46749cc |
|
BLAKE2b-256 | 7319099ba9ccbffcfd4a67de2a362b5fdaa53a3159f73c56d091d8b900ca9f77 |